JP2012517774A5

JP2012517774A5 -

Info

Publication number: JP2012517774A5
Application number: JP2011550203A
Authority: JP
Filing date: 2010-02-10
Publication date: 2013-04-11

Claims

A method for detecting block noise in an image,
Calculating a difference amount of pixel values of adjacent pixels in the image;
Removing a set of difference amounts from the difference amounts caused by features in the image;
Applying a median filter to the image to generate an output;
Subtracting the output from the difference amount;
Performing edge detection in the image to detect edge pixel positions;
A method comprising the steps of:

The difference amount includes a horizontal difference amount of horizontally adjacent pixels and a vertical difference amount of vertically adjacent pixels,
Averaging the horizontal difference amount in a vertical direction to generate a first vector of the average value of the horizontal difference amounts of the width of the image;
Averaging the vertical difference amount in a horizontal direction to generate a second vector of the average value of the vertical difference amount of the height of the image;
The method of claim 1 further comprising:

Further plotting the first vector on a graph against pixel locations along the width of the image;
The method according to claim 2.

Plotting the second vector in a graph against the pixel locations along the height of the image;
The method according to claim 2.

Further comprising removing a DC level from the average value of the vertical difference amount or the average value of the horizontal difference amount;
The method according to claim 2.

The DC level is removed by applying a median filter to the average value and subtracting the result from the average value;
6. The method of claim 5, wherein:

Further comprising suppressing the values of the first and second vectors near the edge of the image;
The method according to claim 2.

Obtaining a normalized sum of the differences for a set of distance size candidates and a set of offset candidates;
Determining the block noise distance size and offset using the normalized sum maximum;
The method of claim 2, further comprising:

Further using the distance size and offset to determine a nominal block size and a starting boundary block position;
The method according to claim 8, wherein:

Further comprising determining the sum of three normalized difference amounts;
For the first sum of the three normalized sums, use the nominal position and the nominal position −1 to calculate the sum, and select the position that yields the greater of the sums;
For the second sum of the three normalized sums, use the nominal position and the nominal position +1 to calculate the sum, and select the position that yields the greater of the sums;
In a third sum of the three normalized sums, use the nominal position to calculate a sum and select the nominal position;
The method according to claim 8, wherein:

Identifying distance size candidates and offset values that produce a maximally normalized difference amount; and
The maximally normalized sum is compared with the sum calculated for the other two offset values of the distance size candidate to determine the block size of the block noise, and the nominal block size and the starting boundary position are Using to select the final block size and offset value;
The method of claim 10, further comprising:

Further comprising identifying a next boundary block location;
The next boundary block position is
Incrementing the starting block position by the nominal block size;
Identifying the position of the maximum or near-maximum difference amount value in a window around the next boundary block position;
Selecting the position at or near the maximum as the next block boundary position;
Sought by,
The method of claim 9.

Further comprising generating a block boundary map using the plurality of next boundary block positions;
The method according to claim 12.

Further comprising determining whether the image is scaled using the difference amount;
The method according to claim 1.

The horizontally and vertically adjacent pixels are one pixel difference,
The method according to claim 1.

Further comprising the step of attenuating the set of difference amounts by identifying the set of difference amounts equal to or greater than a threshold and setting the set of difference amounts to the threshold
The method according to claim 1.

Subtracting the output of the median filter from the difference amount further comprises rounding a value less than zero to zero.
The method according to claim 1.

The edge detection is performed using a Sobel filter,
The method according to claim 1.

The Sobel filter is 3 × 3,
The method according to claim 18, wherein:

The step of suppressing the edge difference amount at the edge pixel position further includes the step of setting the edge difference amount at the position of the edge pixel position whose corresponding gradient magnitude is equal to or greater than a gradient threshold to zero.
The method according to claim 2.

The difference amount is calculated with respect to luminance and saturation.
The method according to claim 2.

Generating a saturation block boundary map using the difference amount calculated for the saturation, and generating a luminance block boundary map using the difference amount calculated for the luminance;
The method according to claim 21, wherein:

Further comprising determining a set of noise characteristics of the block noise using the normalized sum and determining a set of image characteristics of the image using difference data.
The method according to claim 8, wherein:

The set of noise characteristics includes block noise intensity and block size, and the set of image characteristics includes a detailed amount in the image;
24. The method of claim 23.

Further comprising determining a gain value of a block noise filter using the set of noise characteristics and image characteristics;
24. The method of claim 23.

A method of configuring a block noise filter to filter block noise in an image, comprising:
Calculating a difference amount of pixel values of adjacent pixels in the image, including a horizontal difference amount of horizontally adjacent pixels and a vertical difference amount of vertically adjacent pixels;
Determining a set of noise characteristics of the block noise using the normalized sum of the difference amounts;
Obtaining a set of image characteristics of the image using difference amount data;
Configuring the input to the block noise filter using the set of noise characteristics and image characteristics;
A method comprising the steps of:

The input to the block noise filter includes a gain value having a vertical filter gain and a horizontal filter gain;
27. A method according to claim 26.

The set of noise characteristics includes block noise intensity and block size;
The set of image characteristics includes a detailed amount in the image;
27. A method according to claim 26.

The gain value includes a horizontal filter gain and a vertical filter gain,
27. A method according to claim 26.

Further comprising increasing the gain value of the block noise filter when the intensity of the block noise increases;
28. The method of claim 27.

Further comprising increasing the gain value of the block noise filter when the amount of detail in the image is small;
28. The method of claim 27.

Further comprising increasing the gain value of the block noise filter when block noise is high,
28. The method of claim 27.

The input to the block noise filter further comprises an edge gradient multiplier having a horizontal edge multiplier and a vertical edge multiplier;
27. A method according to claim 26.

Further comprising decreasing the edge gradient multiplier when the intensity of block noise increases;
34. The method of claim 33.

Further increasing the edge gradient multiplier when the amount of detail in the image is large;
34. The method of claim 33.

Generating a block boundary map of block noise boundary locations in the image, wherein the block boundary map includes vertical and horizontal boundaries;
27. A method according to claim 26.

Configuring the block noise filter with the block boundary map, the gain value, and the edge gradient multiplier;
Generating a filtered image by applying the block noise filter to the image;
The method of claim 26, further comprising:

Further comprising edge blending the filtered image;
38. The method of claim 37.

Further comprising edge smoothing the filtered image.
38. The method of claim 37.

A method of applying a block noise filter to an image,
Selecting a gain setting including a vertical gain coefficient of the block noise filter based on block noise characteristics in the image;
Calculating a set of coefficients for a vertical luminance filter using the vertical gain coefficients;
Calculating a set of coefficients for a vertical saturation filter using the vertical gain coefficients;
Filtering the image along a vertical dimension using the coefficients of the vertical filter and saturation filter to generate a vertically filtered image;
A method comprising the steps of:

The gain setting includes a horizontal gain factor;
Calculating a set of coefficients for a horizontal luminance filter using the horizontal gain coefficients;
Calculating a set of coefficients for a horizontal saturation filter using the horizontal gain coefficient;
Filtering the image along a horizontal dimension using the coefficients of the horizontal filter and saturation filter to generate a horizontally filtered image;
41. The method of claim 40, further comprising:

Generating a first mixed image by mixing the image with the vertical filtered image;
Generating a second mixed image by mixing the vertical filtered image and the horizontal filtered image;
42. The method of claim 41, further comprising:

Further comprising edge smoothing the mixed image;
43. The method of claim 42.

Means for calculating a difference amount of pixel values of adjacent pixels in the image;
Means for excluding a set of difference amounts from the difference amounts caused by features in the image;
Means for determining the sum of the normalized difference amounts of a set of distance size candidates and a set of offset candidates;
Means for calculating and obtaining an average of the row and column of the difference value;
Means for determining a nominal block size and starting boundary block location using the distance size and offset;
Means for identifying a next boundary block position determined by incrementing the starting block position using the nominal block size;
Means for determining a distance size and offset of the block noise using the normalized sum maximum;
Means for generating a block boundary map using a plurality of next boundary block positions;
A system comprising:

An optical disk reader that reads the optical disk during operation;
A memory coupled to the optical disc reader;
A graphics processor coupled to the memory;
The graphics processor executes an algorithm that applies a block noise filter to the image, the algorithm comprising:
Selecting a gain setting including a vertical gain coefficient and a horizontal gain coefficient of the block noise filter based on block noise characteristics in the image;
Calculating a coefficient of a vertical luminance filter using the vertical gain coefficient;
Calculating a coefficient of a vertical saturation filter using the vertical gain coefficient;
Using the coefficients of the vertical filter and saturation filter to filter the image along a vertical dimension to produce a vertically filtered image;
Calculating a coefficient of a horizontal luminance filter using the horizontal gain coefficient;
Calculating a coefficient of a horizontal saturation filter using the horizontal gain coefficient;
Using the coefficients of the horizontal filter and saturation filter to filter the vertically filtered image along a horizontal dimension to generate a horizontally filtered image;
Generating a first mixed image by mixing the vertical filtered image and the image;
Generating a second mixed image by mixing the vertical filtered image and the horizontal filtered image;
Executed by the
An optical disc device characterized by the above.

The optical disc reader is a Blu-ray disc reader;
46. The apparatus of claim 45.

The optical disc reader is a digital versatile disc ( DVD ) reader;
46. The apparatus of claim 45.

The optical disc reader is a high definition DVD ( HD-DVD ) reader;
46. The apparatus of claim 45 .